Abstract: Annals of Marine Science Research

Abstract:
Desertification is a crucial environmental problem in the Algerian steppe. Land degradation is a reduction and loss of land productivity due to natural processes, climate change and human activities. The sandy lands in the zone of Naâma constitute a region based on important pastures in west of Algeria. It is therefore increasingly necessary to monitor the situation to better understand the desertification processes and to fight against this phenomenon. Climate, soils, vegetation and land use play an important role in desertification. Therefore, the land surface tem perature and a monitoring indicator that highlight the hottest areas as well as their delineation are important to characterize the phenomenon of desertification. In this study, we analyzed desertification since 2008 using remote sensing technology. In order to recognize desertification through land-use changes, our research was aimed at ex tracting classification categories representing the state of the land, which we subdivided into normal classification categories of vegetation groups. To this end, we have classified land use by combining several spectral indices in particular, which can be calculated from satellite data on each Land sat satellite spectral band, to construct multi band input data for a supervised classification approach based on a support vector (SVM). By applying this method to Land sat archival imagery in 2008, 2011, 2014 and 2017, we produced land use maps for the four periods. Using GIS, we integrated climatic parameters with rainfall and land surface temperature combined with land use land cover and slope, in this step; we adopted proposals from expert judgments to determine weights in order to assign weights to each parameter. This allowed us to clarify the situation with regard to desertification, to classify the study area in areas vulnerable to desertification and to determine the spatiotemporal changes during the 10 years.